meridian-scanning photometer, coherent hf radar, and ......coherent hf radar, a meridian-scanning...

Meridian-scanning photometer, coherent HF radar,and magnetometer observations of the cusp: a case study

S. E. Milan1, M. Lester1, S. W. H. Cowley1, J. Moen2, P. E. Sandholt3, C. J. Owen4

1Department of Physics and Astronomy, University of Leicester, Leicester LE1 7RH, UK2UNIS, P.O. Box 156, N-9170 Longyearbyen, Norway3Department of Physics, University of Oslo, P O Box 1048 Blindern, N-0316 Oslo, Norway4Queen Mary and West®eld College, University of London, London E1 4NS, UK

Received: 29 September 1997 / Revised: 25 August 1998 /Accepted: 4 September 1998

Abstract. The dynamics of the cusp region and post-noon sector for an interval of predominantly IMF By, Bz< 0 nT are studied with the CUTLASS Finlandcoherent HF radar, a meridian-scanning photometerlocated at Ny AÊ lesund, Svalbard, and a meridionalnetwork of magnetometers. The scanning mode of theradar is such that one beam is sampled every 14 s, and a30° azimuthal sweep is completed every 2 minutes, all at15 km range resolution. Both the radar backscatter andred line (630 nm) optical observations are closely co-located, especially at their equatorward boundary. Theoptical and radar aurora reveal three dierent behav-iours which can interchange on the scale of minutes, andwhich are believed to be related to the dynamic nature ofenergy and momentum transfer from the solar wind tothe magnetosphere through transient dayside reconnec-tion. Two interpretations of the observations arepresented, based upon the assumed location of theopen/closed ®eld line boundary (OCFLB). In the ®rst,the OCFLB is co-located with equatorward boundary ofthe optical and radar aurora, placing most of theobservations on open ®eld lines. In the second, theobserved aurora are interpreted as the ionosphericfootprint of the region 1 current system, and theOCFLB is placed near the poleward edge of the radarbackscatter and visible aurora; in this interpretation,most of the observations are placed on closed ®eld lines,though transient brightenings of the optical auroraoccur on open ®eld lines. The observations reveal severaltransient features, including poleward and equatorwardsteps in the observed boundaries, ``braiding'' of thebackscatter power, and 2 minute quasi-periodic en-hancements of the plasma drift and optical intensity,predominantly on closed ®eld lines.

Key words. Ionosphere (auroral ionosphere; plasmaconvection) á Magnetospheric physics (magnetopause,cusp, and boundary layers).

1 Introduction

The cusp (see the review papers of Lockwood (1991) andSmith and Lockwood (1996), and references therein) isthat region where entry of magnetosheath plasma intothe magnetosphere, and hence the transport of mass,energy, and momentum from the solar wind into thenear-Earth environment, is facilitated by reconnectionat the dayside magnetopause. While the relative impor-tance of quasi-continuous reconnection or more tran-sient and intermittent ``¯ux transfer events'' (Russell andElphic, 1978; Haerendel et al., 1978; Paschmann et al.,1982; Saunders et al., 1984; Lockwood, 1991) is as yetunclear, there is a considerable body of evidence tosuggest that the auroral signature of the cusp and theionospheric convection excited by dayside reconnectionare pulsed (e.g. Feldstein and Starkov, 1967; Vorobjevet al., 1975; Horwitz and Akasofu, 1977; van Eykenet al., 1984; Goertz et al., 1985; Lockwood et al., 1989a,1989b, 1993b, 1993c; Sandholt et al., 1985, 1986, 1989;Sandholt, 1988; Elphic et al., 1990; Fasel et al., 1992;Pinnock et al., 1993, 1995; Moen et al, 1995, 1996;éieroset et al., 1996). Much theoretical work has beenpublished on the ionospheric convection response totransient dayside reconnection (e.g. Southwood, 1985,1987; Siscoe and Huang, 1985; Lee, 1986; Cowley, 1986;Southwood et al., 1988; Scholer, 1988; Wei and Lee,1990; Lockwood et al., 1990; Cowley et al., 1991;Cowley and Lockwood, 1992; Cowley et al., 1992;Lockwood and Cowley, 1992; Rodger and Pinnock,1997). However, the dynamics of the cusp region are stilllargely unclear.

The present study examines the ionospheric convec-tion ¯ow measured by the CUTLASS Finland HF radarin the vicinity of the cusp aurora observed by ameridian-scanning photometer located at Ny AÊ lesund,Svalbard. Supporting observations are provided byIMAGE magnetometer stations from northern Scandi-navia and interplanetary magnetic ®eld measurementsCorrespondence to: S. E. Milan

Ann. Geophysicae 17, 159±172 (1999) Ó EGS ± Springer-Verlag 1999

from the WIND spacecraft. This study is interestingfrom two perspectives, both in terms of cusp regiondynamics as well as in understanding the processes thatlead to irregularity formation, and hence coherent radarbackscatter, at the ionospheric footprint of the cusp.

2 Experimental arrangement

The SuperDARN coherent HF radars (Greenwald et al.,1995) employ backscatter from high latitude ®eld-aligned ionospheric plasma density irregularities (radaraurora) as tracers of the bulk plasma motion under thein¯uence of the convection electric ®eld, and hence as adiagnostic tool for the investigation of large-scalemagnetospheric-ionospheric coupling. The Co-operativeUK Twin Located Auroral Sounding System (CUT-LASS) forms the eastern-most pair of radars ofSuperDARN, of which the Finland radar, located atHankasalmi (62.3°N, 26.6°E), is the focus of the presentstudy.

The CUTLASS Finland radar usually sounds se-quentially along 16 beams separated by approximately3.2° of azimuth, distributed symmetrically about theradar boresite of )12° (i.e. west of geographic north). Inthe present study, the radar scanned through beams 12to 5 in descending order, returning to beam 9 betweeneach (i.e. beams 12, 9, 11, 9, 10, 9, 9, 9, 8, 9, 7, 9, 6, 9, 5,9), with a dwell time of 7 s, producing a ®eld-of-viewmap of backscatter every 2 minutes. Beam 9 had asampling time resolution of 14 s. 75 range gates weresampled for each beam, with a pulse length of 100 ls,

corresponding to a gate length of 15 km, and a lag to the®rst gate of 8700 ls (1305 km). Fig. 1 illustrates thelocation of the ®eld-of-view of the CUTLASS Finlandradar for this experiment. The statistical location of theauroral oval (Feldstein and Starkov, 1967; Holzworthand Meng, 1975) at 0930 UT is also shown in Fig. 1 forquiet geomagnetic conditions (Kp = 1). A 7 pulsescheme is transmitted, and analysis of the auto-correla-tion function (ACF) of the returned signals yieldsbackscatter Doppler spectra, from which the spectralcharacteristics of power, line-of-sight Doppler velocityand spectral width can be derived. The Doppler velocitygives an estimate of the radar line-of-sight component ofthe plasma convection velocity.

The Meridian-Scanning Photometer (MSP) locatedat Ny AÊ lesund (78.9°N, 11.9°E), Svalbard, scansroughly along the magnetic meridian, 35° west ofgeographic north, to 10° above the northern andsouthern horizons, with a scan period of 18 s. Themeridian closely follows beam 9 of the CUTLASSFinland radar (see Fig. 1). The line-of-sight intensities of630 nm (red line) auroral emissions are measured,corresponding to transitions from the 1D metastablestate of atomic oxygen (OI). These auroral emissionscan have an altitude of peak emission intensity between250 km and 550 km, excited by electrons of energies oforder 100 eV. In the present study, an emission altitudeof 250 km is assumed.

Seven IMAGE magnetometer stations are alsoincluded in the study. The stations NAL, HOR, BJN,SOR, MAS, and MUO form a meridional chain, and thestation HOP is located between HOR and BJN in

Fig. 1. The locations of the ®eld-of-view of theCUTLASS Finland radar (beam 9 is explicitly in-dicated) and the meridian of the Ny AÊ lesundMSP, projected to 250 km altitude. Numbers tothe right of the MSP meridian indicate thelocation of the projection of zenith angles 0°, 65°and 80°. The locations of the IMAGE magne-tometer sites at NAL, HOR, HOP, BJN, SOR,MAS, and MUO are indicated. Also illustratedfor reference is the statistical location of the quiet(Kp = 1) auroral oval at 0930 UT

160 S. E. Milan et al.: Meridian-scanning photometer, coherent HF radar, and magnetometer observations of the cusp

latitude and approximately 250 km to the east of thechain (see Fig. 1).

3 Observations

During the interval considered in the present study, 0840UT to 1030 UT on 17 December 1995, the Ny AÊ lesundMSP, the CUTLASS Finland radar, and the IMAGEmagnetometers were located near local noon and in theearly post-noon sector, and provided information re-garding the ionospheric footprint of the magnetosphericcusp. Solar wind and interplanetary magnetic ®eld(IMF) observations from the WIND spacecraft providea context within which the behaviour of the ionosphericcusp region can be investigated.

3.1 Solar wind and IMF observations

During the interval of study the WIND spacecraft waslocated upstream of the magnetosphere at (XGSM,YGSM, ZGSM) � (59 Re, 32 Re, 0 Re). WIND measure-ments of the solar wind dynamic pressure and the threeGSM components of the IMF, Bx, By, and Bz, are

illustrated in Fig. 2 for the interval 0824 UT to 1014UT. A rotation of the IMF was observed near 0845 UT,associated with a 40% increase in the solar winddynamic pressure. Subsequently, the southward compo-nent of the IMF, Bz, remained predominantly negative,though decreasing in magnitude; By remained negativefor the rest of the interval of study.

An approximate solar wind velocity of 420 km s)1

measured by the spacecraft during this interval gives adelay of some 12 minutes between the observation ofIMF features by WIND and their incidence on theEarth's bow-shock. Various assumptions can be maderegarding the subsequent travel-time of the shockedsolar wind across the magnetosheath and the AlfveÂ ntravel-time for occurrences at the magnetopause to betransmitted to the ionospheric footprint of the cusp,giving an overall estimated delay of some 16 to 24minutes.

3.2 Radar and MSP observations

The MSP observations of the red line intensity for thestudy interval are illustrated in Fig. 3, as a function ofzenith angle and time, from 0800 UT to 1100 UT.

Fig. 2. Time series of the solar wind dynamic pressureand the GSM components of the IMF, Bx, By, andBz, observed by the WIND spacecraft for 0824 UT to1014 UT, 17 December 1995

S. E. Milan et al.: Meridian-scanning photometer, coherent HF radar, and magnetometer observations of the cusp 161

Throughout the interval, a persistent or continuous arcis observed near 45° zenith angle, though this movespoleward after 1000 UT. The intensity of this arc,initially 1.5 kR, increases to 2.5 kR near 0900 UT andagain to 5 kR after 0915 UT, after which the intensityvaries considerably between 2.5 kR and 5 kR. Inaddition to the continuous arc, very faint poleward-moving forms are observed between 0800 UT and 0900UT; three such features are indicated by arrows in panela of Fig. 3. After 0915 UT, these forms appear more as

stationary auroral brightenings near the poleward edgeof the main optical arc; three such features are indicatedin panel b of Fig. 3.

The CUTLASS Finland radar observations frombeam 9, the high time-resolution beam, for the studyinterval are presented in Fig. 4. The top three panelsindicate the radar backscatter power (signal-to-noiseratio), line-of-sight velocity and spectral width. Positive(negative) line-of-sight velocities are towards (awayfrom) the radar. Vertical power drop-outs, especially

Fig. 3. The Ny AÊ lesund MSP630 nm (red line) optical emis-sion intensity as a function ofzenith angle ()90° to the north,90° to the south) and UT, be-tween 08 UT and 11 UT, 17December 1995. Vertical tickmarks are equivalent to 1.5 kRemission intensity. Arrows inpanel a indicate the times ofthree poleward-moving auroralforms. Arrows in panel b indicatethe times of three transientbrightenings of the aurora pole-ward of the continuous arc (seeFig. 4 for the identi®cation ofother transient brightenings)


after 1000 UT, are associated with HF interference. Lowpower, low velocity and low spectral width backscatterobserved at the very nearest ranges between 0840 UT to0915 UT and 1000 UT to 1020 UT is ground backscatterand can be ignored. Similar backscatter observed above79°N between 0925 UT and 1020 UT is shown byinterferometric techniques (see Milan et al., 1997) tohave a dierent elevation angle distribution from therest of the observations and is almost certainly groundbackscatter also; this backscatter region will not beconsidered further. The bottom panel of Fig. 4 illus-trates the red line optical emission intensity along theMSP meridian (i.e. the same data as presented in Fig. 3),projected to an emission altitude of 250 km. Superim-posed on the radar observations is the latitude of thepeak red line emission intensity and the latitudes atwhich the red line emission intensity falls to 1.5 kRpoleward and equatorward of the peak. Vertical dashedlines in Fig. 4 indicate times at which the behaviour ofthe observations appears to change markedly. Theintervals between dashed lines have been numbered Ito VII to aid description in the following discussion. Theoccurrences of transient brightenings of the opticalaurora poleward of the continuous arc are marked in thebottom panel by white arrows.

During interval I only sporadic radar backscatter andlow intensity optical emission is observed. The amountof backscatter observed by the radar and the intensity ofthe optical emission increases signi®cantly at the start ofinterval II, 0900 UT; signi®cant radar and opticalaurora continue to be observed until after 1100 UT.During this period there is a general poleward migrationof the aurora, from near 74° to 77° geomagneticlatitude. In general there is excellent correspondencebetween the locations of the radar and optical cuspaurora, in particular at their equatorward boundary,and especially during intervals II to V. To illustrate thedegree of co-location of the equatorward boundaries ofthe optical emission and radar backscatter, Fig. 5presents the latitudinal variation in the equatorwardiso-contours of 1.5 kR red line emission intensity and 10dB backscatter power, between 0855 UT and 1000 UT.Even during intervals when the contours are separatedin latitude (in general by no more than 0.5°) theirmotions are in tandem. A correlation coecient of 0.90is found between the locations of the two boundaries, asillustrated in the inset panel of Fig. 5, in which thedotted line is a least squares ®t to the distribution, with agradient of 1.01. The separation in latitude of thecontours appears to increase somewhat throughout theinterval presented, perhaps as a consequence of agradual variation in the peak emission altitude of theaurora or of a change in the propagation environment ofthe radar.

In the period of most interest to the present study,intervals II to VI, the observations cycle through twomodes of behaviour, which will hereafter be refered to asstates 1 and 2. State 1, intervals III and V, is charac-terised by high red line optical intensities, a peakintensity above 3.5 kR, and radar line-of-sight velocitieswhich are positive at lower latitudes and negative at

higher latitudes. State 2, intervals II, IV and VI, ischaracterised by lower red line optical intensity, a peakintensity in general below 3.5 kR, and radar line-of-sightvelocities which are negative at all latitudes. These twostates will be described in more detail below.

3.2.1 State 1 (intervals III and V)

The high time resolution radar observations (Fig. 4)indicate that in state 1 the radar backscatter is containedalmost entirely within the higher and lower latitude 1.5kR red line contours. The radar line-of-sight velocitiesare structured in latitude, generally positive at lowerlatitudes and negative at higher latitudes. Spatial mapsof the line-of-sight velocity from these intervals indicatethat the ¯ow is predominantly azimuthal during theseintervals, and that a ¯ow reversal is present, ¯ow beingdirected westward at lower latitudes and eastward athigher latitudes. Fluctuations in the sign of the line-of-sight velocities observed in Fig. 4 during these intervalsare caused by small variations in the angle of ¯owrelative to the nearly polewards-directed beam 9.

The onset of state 1 ¯ow behaviour, the beginning ofintervals III and V, is accompanied by the appearance ofa new peak in the backscatter power at a latitude belowthat of the pre-existing peak (see 0914 UT and 0950 UT,Fig. 4). This occurs also at 0921 UT, during on-goingstate 1 behaviour. This feature will hereafter be termed``braiding'' as in some cases the pre-existing peak in thebackscatter power migrates poleward prior to theappearance of the second peak, as at 0914 UT and0921 UT. During state 1, when two backscatter intensitypeaks are present, the central minimum appears at thelatitude of the ¯ow reversal, also close to the latitude ofthe peak in the red line optical intensity. A similarlatitudinal structure is observed in the spectral widthmeasurements during state 1, higher spectral widthsbeing co-located with the peaks in backscatter intensity(see esp. interval III).

3.2.2 State 2 (intervals II and IV)

During state 2, the line-of-sight velocities in the hightime resolution observations become exclusively nega-tive. Spatial maps from these intervals suggest that theplasma drift continues to display a general azimuthal¯ow, eastward and westward at higher and lowerlatitudes, respectively. However, the state 2 plasma driftvelocity is signi®cantly structured in time, with recurrentvelocity enhancements moving rapidly poleward (seeesp. interval IV, Fig. 4). These velocity enhancementsare quasi-periodic in nature, with a recurrence periodcomparable with the radar scan period and hence theradar scan has insucient temporal resolution toproperly resolve these features. Transient enhancementsare also present in the red line emission intensitiesduring state 2 (esp. intervals IV and VI). To investigatethe relationship between these optical and plasma drift


transients, Fig. 6 presents a more detailed comparisonof the velocity and optical measurements. The top panelof Fig. 6 illustrates the line-of-sight radar velocities, on ascale from 0 m s)1 to )1000 m s)1 (i.e. away from theradar only). The bottom panel indicates the opticalemission intensity of the red line, projected to anassumed emission altitude of 250 km. Superimposedon both panels are 10 diagonal dashed lines, indicating a®t-by-eye to the poleward-moving features, based pri-marily on the optical observations. In most of the 10cases there is an excellent agreement between thelocation and motion of the poleward-moving opticalfeatures and the poleward-moving velocity enhance-ments. The lower-latitude boundary of the velocityenhancements matches well the lowest latitudes at whichthe intensi®cations in the red line emission are observed.However, the transients are observed to continue tohigher latitudes in the radar data than in the opticalmeasurements. The velocity and optical intensity en-hancement features have a lifetime and recurrenceperiod of approximately 2 minutes.

3.3 Magnetometer observations

Fig. 7 illustrates magnetograms (X and Y components)for the study interval from the 7 IMAGE stations. The

magnetograms for the interval 0855 UT to 0910 UT (i.e.spanning the transition from interval I to interval II)from the stations NAL, HOR, HOP, and BJN indicate apositive (negative) bay in the X component of stationsequatorward (poleward) of HOP and a negative-then-positive bipolar signature in the Y component of all thestations. Such signatures have been identi®ed with``travelling convection vortices'' or TCVs (Friis-Chris-tensen et al., 1988), but this signature is also consistentwith a variation in the intensity of a FAC sheet locatednear the latitude of HOP (approximately 76°N). Thetransition from interval III to interval IV (approximate-ly 0926 UT to 0930 UT) is again marked a similar,though less intense, magnetic signature, the centre beinglocated this time between the latitudes of NAL andHOR (78°N).

4 Discussion

4.1 General morphology and IMF dependenceof the dayside aurora

The degree of co-location of the optical red line andradar aurora is impressive (see also Rodger et al., 1995),especially at their equatorward boundary. This co-location suggests that the precipitation responsible forthe excitation of the optical aurora is also a majorsource of the energy leading to the generation of theionospheric irregularities from which the radar scatters.The source of this precipitation, however, has to beconsidered carefully.

Red line aurora are often observed poleward of thedayside auroral oval (Whalen et al., 1971; Murphreeet al., 1980; Elphinstone et al., 1992), such emissionsbeing excited by low energy (of order 100 eV) electrons ±consistent with the energy characteristics of unacceler-ated magnetosheath plasma of cusp/cleft origin ± onnewly-opened ®eld lines (Lockwood et al., 1993a;Sandholt et al., 1996). Hence, the equatorwards boun-dary of the red line aurora could be related to the

Fig. 5. The locations of the equatorwardboundaries of the red line optical emission(solid line) and radar backscatter (dashedline) between 0855 UT and 1000 UT, asdetermined from the 1.5 kR and 10 dB iso-contours, respectively. The dotted line indi-cates an approximate ®t-by-eye to the motionof these boundaries. Also included is thelocation of the peak in the red line opticalemission (dot-dashed line). Arrows indicatethe approximate times of the occurrence ofequatorward steps of the radar and opticalaurora. The inset panel indicates the corre-spondence between the latitudes of theequatorward boundaries of the radar andoptical aurora, for the period illustrated

Fig. 4. The backscatter power (top panel), line-of-sight velocity(second panel), and spectral width (third panel) observations alongbeam 9 of the CUTLASS Finland radar for the period 0840 UT to1030 UT. Positive (negative) velocities indicate motion towards (awayfrom) the radar. Also illustrated for the same period is the 630 nm(red line) optical emission intensity observed by the Ny AÊ lesundMSP,projected to an assumed emission altitude of 250 km. White arrowsindicate the times of transient brightenings of the aurora poleward ofthe continuous arc. Superimposed on the radar data are, in order ofincreasing latitude, the locations of the equatorwards 1.5 kR iso-contour, the peak intensity, and the polewards 1.5 kR iso-contour ofthe red line emission. The latitude range of the panels, 75°N to 81°N,corresponds to the geomagnetic latitude range 72.6° to 78.5°CGMLAT

b


location of the open/closed ®eld line boundary orOCFLB.

However, other sources of precipitation are alsopresent in the dayside high latitude ionosphere, espe-cially in the pre- and post-noon sectors, for instance thelow latitude boundary layer (LLBL) and boundaryplasma sheet (BPS) precipitation on closed ®eld lineswhich constitute the region 1 ®eld-aligned currentsystem (Potemra, 1994). In the post-noon sector, wheremost of the present observations are made, the region 1FAC is directed upwards. This LLBL/BPS precipitation,as identi®ed by DMSP particle measurements, gives riseto ``cleft aurora'', a persistant or continuous band of redline emission, associated with sunward plasma drift(Sandholt et al., 1993); such aurora were labeled ``type7'' by Sandholt et al. (1998). Poleward of the cleft aurorais observed mantle precipitation, associated with anti-sunward plasma drift (Sandholt et al., 1993). Thismantle precipitation occurs on newly-opened ®eld lineswhich are a few minutes old, and hence which haveconvected some distance from the cusp region. It is thisprecipitation which is associated with auroral brighten-ings poleward of the cleft aurora. Hence an alternativeinterpretation of the present observations, certainly after0915 UT, is of continuous cleft aurora on closed ®eldlines at the lower latitudes, and transient auroralbrightenings associated with precipitation on open ®eldlines at higher latitudes; three such auroral brighteningsare indicated in panel b of Fig. 3. The OCFLB is, then,

placed near the poleward edge of the continuous red lineaurora, but equatorward of the transient brightenings.

It is possible that both of the above scenarios areobserved during the present interval. Prior to 0900 UT,interval I, the red line aurora and radar backscatter areperhaps associated with the cusp-proper, in a con®gu-ration known as the ``midday gap'' ± low red line opticalintensity, sometimes associated with poleward-movingauroral forms, as indicated in panel a of Fig. 3. Afterthis time, at the beginning of interval II, the appearanceof the region 1 current system equatorward of the cusp/mantle precipitation gives rise to an increase in theintensity and latitudinal extent of the optical and radaraurora. The change in the auroral con®guration at thebeginning of interval II appears to be a consequence ofthe rotation of the IMF direction observed by theWIND spacecraft near 0840 UT. The Bz component ofthe IMF remains southward or near-zero throughoutthe interval, but there is a swap in the east-westcomponent from By > 0 nT to By < 0 nT. This changeimpinges on the dayside magnetopause after a solarwind delay from the WIND location of some 16 to 24minutes, i.e. near 0900 UT. The ionospheric footprint ofthe dayside merging line then moves from the post- topre-noon sector under the in¯uence of IMF By, and apost-noon auroral con®guration is established overheadthe radar and MSP.

Also at this time, a sudden compression of themagnetosphere is expected, due to step in the solar wind

Fig. 6. Radar and MSP obser-vations of the period 0923 UT to0947 UT (end of interval III andinterval IV). (a) Line-of-sightplasma drift velocities alongbeam 9 (note: the scale involvesnegative velocities only). (b) Redline optical emission intensitiesprojected to 250 km. The dashedlines in both panels are ®ts-by-eye to the optical poleward-moving optical transients and theradar ¯ow bursts


dynamic pressure observed at 0843 UT, near the time ofthe IMF rotation. Such pressure changes are known togive rise to ``travelling convection vortices'' or TCVs(e.g. Friis-Christensen et al., 1988), similar to thatobserved in the magnetograms between 0855 UT and0910 UT (see Sect. 3.3 and Fig. 7). It is just as

reasonable to suggest, however, that the observedmagnetic de¯ections are associated with the motionand intensi®cation of the FAC systems associated withthe change in IMF By. As no other signi®cant pressurevariations were observed during the present study, allsubsequent variations in the auroral behaviour must be

Fig. 7. X and Y componentmagnetograms from the IM-AGE stations NAL, HOR,HOP, BJN, SOR, MAS, andMUO, for the period 0840 UT to1030 UT on 17 December 1995


produced by variations in the IMF, including the TCV-like signature observed between intervals III and IV (seeSect. 3.3).

The plasma ¯ow and optical behaviour observedafter 0900 UT appears to interchange between twocharacteristic states. In both a ¯ow shear is evident nearthe latitude of the peak of the optical intensity.However, in state 2 there appears to be a signi®cantpoleward ¯ow component superimposed on this ¯owreversal pattern. Possibly, the switch between the twostates is due to the motion overhead of the convectionpattern as a whole, under the in¯uence of the orientationof the IMF. In this case, the two states would, perhaps,represent observations from two dierent points withinthe convection pattern, in state 2 a view near the throatof the convection pattern where polewards ¯ow isexpected, and in state 1 a view further around the duskconvection cell where a ¯ow reversal is expected.However, a brief examination of the IMF clock angledoes not reveal any obvious precursors of the inter-change between auroral states. Unfortunately, a moredetailed analysis of the cusp auroral response to changesin the IMF is not possible due to the uncertainty in thedelay between WIND observations of the IMF and theirimpact on the dayside ionosphere.

4.2 Motions of the open/closed ®eld line boundary

A knowledge of the location and motion of theionospheric footprint of the open/closed ®eld lineboundary is important for magnetospheric studies. Ingeneral the OCFLB is ``adiaroic'' (Siscoe and Huang,1985) and there is no plasma ¯ow across this boundary.During periods of dayside reconnection, previouslyclosed ®eld lines are opened to form a region ofnewly-opened ¯ux. At this time a portion of theOCFLB, the ``merging line,'' becomes non-adiaroicand plasma moves relative to the boundary due to thereconnection electric ®eld. Determining the length of themerging line and the magnitude of the reconnectionelectric ®eld allows the contribution of dayside recon-nection to the cross polar cap potential to be estimated(e.g. Baker et al., 1997). In general, the true location of anon-adiaroic OCFLB should lie equatorward of theequatorward boundary of cusp precipitation. This is aconsequence of the non-zero travel-time of the mostenergetic magnetosheath particles from the magneto-pause reconnection site to the ionosphere, during whichtime the newly-opened ¯ux tubes are convecting pole-ward under the in¯uence of the reconnection electric®eld (Rodger and Pinnock, 1997). However, if noreconnection is taking place at the point of observation± as is the case outside of the cusp region proper, i.e.after the start of interval II ± the reconnection electric®eld is zero and the boundary motion and ionospheric¯ow are equal.

In the following discussion, the equatorward boun-dary of the radar and optical aurora will be assumed tobe a proxy for the location of the OCFLB. The OCFLBmay be displaced some distance poleward of this auroral

boundary, esp. if the equatorward portion of the auroralluminosity is the consequence of precipitation in theregion 1 current system on closed ®eld lines. It will beassumed, however, that this oset is approximatelyconstant. Short timescale motions of the dayside boun-dary of the polar cap are assumed to be independent ofsubstorm processes occurring on the nightside. Thegeneral polewards drift of the aurora observed between0900 UT and 1030 UT is probably a consequence of adecrease in the area of the polar cap due to a disparitybetween the dayside and nightside reconnection rates,but the velocity of this motion is small in comparison tothe localised perturbations produced by the dayside cuspphenomena which will now be considered.

The location of the OCFLB, as determined from thecombined radar and optical observations, appears tostep equatorward and retreat poleward in a quasi-periodic manner (Fig. 4 and 5). This motion is possiblyrelated to the transient nature of dayside reconnection.A burst of reconnection ± a ¯ux transfer event or FTE ±appends a new region of open ¯ux to the polar cap (e.g.Cowley and Lockwood, 1992; Cowley et al., 1992;Lockwood and Cowley, 1992), as indicated in Fig. 8(1).During the period of creation of the new open ¯ux, theOCFLB bounding this region ± the merging gap ±expands equatorwards, by perhaps 0.5° of latitude orsome 50 km. Such equatorward motions of the OCFLB,typically with a duration of 2 to 3 minutes, are observedin the present study, for instance at 0900 UT, 0912 UT,0921 UT, 0940 UT and 0949 UT (see Fig. 5, timesmarked with full arrows); there is also some evidence foran equatorward motion of the cusp aurora at 0927 UT(time marked with an open arrow). During intervals ofstate 1 behaviour (intervals III and V) these equator-ward jumps of the OCFLB are accompanied by theappearance of new low latitude peaks in the radarbackscatter ± ``braiding'' ± at 0913 UT, 0921 UT and0950 UT.

Two forces act upon the new open ¯ux tube(Southwood, 1985, 1987; Cowley, 1986), and hence theionospheric plasma at its footprint: one azimuthally dueto the By component of the IMF (Atkinson, 1972;Jùrgensen et al., 1972) and the other poleward due to theantisunward magnetosheath ¯ow (Saunders, 1989).Initially, the poleward component is negligible as thenewly-opened ¯ux tube is draped across the front of themagnetopause and slack, and the azimuthal componentof the ¯ow dominates, as in Fig. 8(2); the sense of Byduring the interval of study produces an eastwardazimuthal motion. After a delay of some 5 to 8 minutesthe slack in the newly-opened ¯ux tube is taken up, andthe poleward velocity component becomes increasinglysigni®cant (Cowley et al., 1992). This poleward motionacts to assimilate the new region of open ¯ux with thepre-existing open ¯ux of the polar cap, producing apoleward retreat of the adiaroic OCFLB, returning thesystem to quasi-equilibrium. This poleward motion ofthe OCFLB has, in the present observations, a meanvelocity of order 125 m s)1. At this velocity, the OCFLB,its location initially perturbed by some 50 km followingthe burst of reconnection, can return to its equilibrium


position in approximately 5 to 10 minutes. Once the FTE¯ux tube is assimilated with the old open ¯ux of the polarcap, Fig. 8(3), the forces acting on it become negligibleand the ¯ow decays. Outside of the moving FTE ¯uxtube footprint, ionospheric ¯ow is produced primarily bydisplacement of the surrounding (incompressible) openand closed ¯ux tubes; it is this displacement that isresponsible for the excitation of convection within thepolar cap and return ¯ow at lower latitudes. The region 1current system on closed ®eld lines equatorward of theOCFLB will be displaced equatorward, and hence itslower latitude boundary will act as a proxy for themotion of the OCFLB; see Fig. 8(2).

The transient equatorward steps of the OCFLB are ingeneral accompanied by the appearance of transientenhancements of the optical luminosity poleward of the

main continuous region 1 current arc, most obvious atthe times marked with arrows in the bottom panel ofFig. 4. These are interpreted as luminosity associatedwith precipitation within the region of newly opened ¯uxas it convects eastwards across the MSP meridian. Theseoptical transients occur within the region of antisunward(eastward) ¯ow observed by the radar.

4.3 Flux transfer events and shorter periodtransient events

As discussed above, many of the observations of thepresent study can be explained in terms of ¯ux transferevents. Magnetopause FTEs ± the magnetic and particlesignature of transient dayside reconnection ± have arecurrence period of 8 minutes (Berchem and Russell,1984; Rijnbeek et al., 1984), though Lockwood andWild (1993) report an inter-FTE interval distributionwhich extends to periods as short as 2 to 3 minutes.Ground-based studies of the optical cusp aurora havereported poleward-moving optical transients ± ``daysideauroral breakups'' ± which propagate poleward fromthe red line aurora with velocities of order 0.5 km s)1 to1 km s)1 (Feldstein and Starkov, 1967; Vorobjev et al.,1975; Horwitz and Akasofu, 1977; Sandholt et al., 1985;Sandholt, 1988; Karlson et al., 1996), associated withmagnetopause FTEs (Sandholt et al., 1986; Elphic et al.,1990). These transients are visible in the green line(557.7 nm) as well as the red, indicating the precipitationof electrons accelerated to keV energies, i.e. not of cusp-proper energy characteristics as de®ned by Newell andMeng (1988). Fasel (1995) determined that these eventshad a mean recurrence period of 7 minutes, though theinter-event interval distribution was highly skewed andhad a mode of 3 minutes. In the present study,observations of exactly this nature are con®ned tointerval I, see panel a of Fig. 3. Previous HF radarstudies of the cusp region have indicated poleward-moving velocity enhancements, major events having arecurrence period of 7 minutes, and these also have beenassociated with FTEs (Pinnock et al., 1995).

In the present study, the observations appear to bemade predominantly to the east of the cusp proper.Equatorward steps of the low latitude boundary of theradar and optical aurora, and transient brightenings ofthe aurora poleward of the main persistent arc areattributed to the eastward motion of regions of newly-opened ¯ux ± each associated with an individual FTE ±from the cusp region and into the radar ®eld-of-view.These observations bear similarities to other studiesconducted in the post-noon sector, such as Moen et al.(1995, 1996) and Provan et al. (1998). Nine transientbrightenings of the aurora are observed within theinterval 0920 UT to 1025 UT, giving an approximatemean recurrence period of just over 8 minutes, close tothe recurrence periods of magnetopause FTEs (Berchamand Russell, 1984; Rijnbeek et al., 1984), of opticaltransients (Fasel, 1995), and of plasma drift velocitybursts (Pinnock et al., 1995). Overall it is estimated thatthe growth and decay of ionospheric convection ¯ow

Fig. 8. A schematic diagram indicating three stages in the assimila-tion of a region of newly-opened ¯ux with the old open ¯ux of thepolar cap (inside semi-circle) for By < 0 nT. Local noon is at the topof the diagram and the approximate location of the MSP meridian isindicated in the post-noon sector. The solid line is the adiaroic open/closed ®eld line boundary. Dashed lines indicate the boundarybetween new and old open ¯ux. The light grey regions in the pre- andpost-noon sectors on closed ®eld lines indicate the region 1 currentlocation. (1) FTE footprint appended to front of polar cap at pre-noon reconnection site. The dot-dashed line represents the merginggap along which reconnection is on-going. This region subsequentlydrifts eastwards. (2) Region of newly-opened ¯ux drifts into radar®eld-of-view and an equatorward step of the equatorward boundaryof the aurora is observed. (3) Newly-opened ¯ux assimilated withpolar cap, OCFLB returned to quasi-equilibrium, drift velocitybecoming negligible


associated with a single burst of reconnection has a timescale of some 10 to 15 minutes (as discussed in Sect 4.2;see also Cowley et al., 1992; Lockwood and Davis,1996). As the recurrence rate of FTEs is of order 8minutes, several ``active'' FTE footprints can exist in thedayside ionosphere simultaneously, giving a ¯ow patternindistinguishable from the traditional two-cell convec-tion pattern (Lockwood, 1991).

Other, shorter-period transients are observed in thepresent study also. The high time resolution beam (beam9) line-of-sight drift velocity of state 2 ¯ow, especiallyduring interval IV, is modulated by poleward-moving¯ow bursts which have a lifetime and recurrence periodof approximately 2 minutes (see Fig. 6 and Sect. 3.2.2).Co-located with these velocity transients, and moving intandem with them, are enhancements of the opticalemission intensity. Such optical transients are observedduring both intervals IV and VI, and are located withinthe continuous low latitude optical arc, interpreted asclosed ®eld lines. The velocity enhancements are ob-served within radar aurora associated with, and pole-ward of, the continuous optical arc, i.e. the velocitytransients extend to higher latitudes than the opticaltransients, perhaps into the open ®eld line region of thepolar cap. These transient events do not appear directlyrelated to the FTE signatures discussed above, thoughthey do bear some similarity with the cusp observationsof Pinnock et al. (1995), i.e. polewards-moving velocityenhancements. The 2 minute recurrence period is shorterthan the timescales usually associated with FTEs. Thesetransients appear highly periodic in nature, suggestingan association with magnetohydrodynamic wave phe-nomena. There is some evidence for wave activity in theX component magnetometer observations throughoutintervals III to VI, especially at the lower latitudestations SOR, MAS, and MUO, located far equator-ward of the cusp region. However, these waves have aperiod of approximately 7 minutes, signi®cantly longerthan that of the transients, indeed commensurate withthat expected for high latitude ®eld line resonances.Pulsations with these longer periods have previouslybeen observed on closed ®eld lines, equatorward ofeastwards propagating patches of backscatter associatedwith FTEs (Walker et al., 1986).

5 Summary

Meridian-scanning photometer, coherent HF radar, andmagnetometer observations of the noon and post-noondayside auroral zone have been combined for an intervalof predominantly IMF By, Bz < 0 nT conditions. In thecusp region proper, the red line optical observations areconsistent with a ``midday gap'' con®guration, withoccasional faint poleward-moving auroral forms, asso-ciated with weak and sporadic radar backscatter. In thepost-noon sector, a much more intense continuous orpersistent auroral arc is present, interpreted as LLBL/BPS precipitation, the upward region 1 current system,on closed ®eld lines. Poleward of this, periodic enhance-ments of the optical intensity are interpreted as the

eastward passage of regions of mantle precipitation on®eld lines newly-opened in the cusp region, to the westof the MSP meridian. These periodic enhancementshave a recurrence period of just over 8 minutes,consistent with the reported mean recurrence period ofmagnetopause ¯ux transfer events. In general, the lowerlatitude boundary of the continuous arc steps equator-ward and then recovers poleward with each passage ofan FTE footprint. This is interpreted as a proxy for themotion of the open/closed ®eld line boundary at higherlatitudes, caused by the perturbation of the closed ®eldlines at lower latitudes. The study reveals excellent co-location of the radar and optical aurora, especially attheir equatorward boundary. The radar plasma driftmeasurements indicate that a convection reversal boun-dary exists between westward (sunward) ¯ow at lowerlatitudes and eastward (antisunward) ¯ow at higherlatitudes, apparently near the peak in intensity of theregion 1 current optical arc, i.e. on closed ®eld lines. Theoptical enhancements associated with FTEs occurwithin the region of antisunward plasma drift.

The post-noon optical and radar aurora appear toexist in two states. In the ®rst, the optical intensity of theregion 1 current arc is high and azimuthal shear ¯owassociated with the convection reversal boundary dom-inates the plasma drift pattern. In the second, thisoptical intensity is lower, and the plasma drift ismodulated by polewards-propagating velocity enhance-ments. Co-located with these velocity bursts are en-hancements of the region 1 current arc optical intensity.These modulations of drift velocity and optical intensityappear highly periodic, reminiscent of magnetohydro-dynamic ULF wave activity, with a recurrence period ofapproximately 2 minutes. State 1 and state 2 auroralbehaviour interchange on timescales of 10 s of minutes,with no obvious precursor in the IMF observations.

The present study has emphasised the power of theHF radar technique for cusp region studies, especiallywhen combined with optical observations. It is clear,however, that greater spatial coverage and a higher timeresolution is essential for future investigations of cuspdynamics.

Acknowledgements. CUTLASS is supported by the Particle Phys-ics and Astronomy Research Council (PPARC), UK, the SwedishInstitute for Space Physics, Uppsala, and the Finnish Meteoro-logical Institute, Helsinki. SEM is supported on PPARC grant no.GR/L00865. The IMAGE magnetometer data employed in thispaper were collected as a German-Finnish-Norwegian-Polishproject conducted by the Geophysical Research Division of theFinnish Meteorological Institute (FMI/GEO). The WIND MFIobservations were provided by R P Lepping of the Laboratory forExtraterrestrial Physics, NASA/Goddard Space Flight Centre,MD, USA.

Topical Editor K. -H. Glassmeier thanks A. Rodger and K.Baker for their help in evaluating this paper.

References

Atkinson, G., Magnetospheric ¯ows and substorms, in Magneto-sphere-Ionosphere Interactions, ed. K Folkestadt, 203, Univ-ersitetsforlaget, Oslo, Norway, 1972.


Baker, K. B., A. S. Rodger, and G. Lu, HF-radar observations ofthe dayside magnetic merging rate: A Geospace EnvironmentModeling boundary layer campaign study, J. Geophys. Res.,102, 9603±9617, 1997.

Berchem, J., and C. T. Russell, Flux transfer events on themagnetopause: Spatial distribution and controlling factors,J. Geophys. Res., 89, 6689, 1984.

Cowley, S. W. H., The impact of recent observations on theoreticalunderstanding of solar wind-magnetosphere interactions,J. Geomagn. Geoelectr., 38, 1223, 1986.

Cowley, S. W. H., J. P. Morelli, and M. Lockwood, Dependence ofconvective ¯ows and particle precipitation in the high-latitudedayside ionosphere on the X and Y components of theinterplanetary magnetic ®eld, J. Geophys. Res., 96, 5557±5564,1991.

Cowley, S. W. H., and M. Lockwood, Excitation and decay of solarwind-driven ¯ows in the magnetosphere-ionosphere system,Ann. Geophys., 10, 103, 1992.

Cowley, S. W. H., J. P. Morelli, M. P. Freeman, M. Lockwood, andM. F. Smith, Excitation and decay of ¯ows in the magneto-sphere-ionosphere system due to magnetic reconnection atthe dayside magnetopause and in the geomagnetic tail, Pro-ceedings of the First International Conference on Substorms(ICS-1), Kiruna, Sweden, March 1992 (ESA SP-335), 117±123,1992.

Elphic, R. C., M. Lockwood, S. W. H. Cowley, and P. E. Sandholt,Signatures of ¯ux transfer events at the dayside magnetopauseand in the ionosphere: Combined ISEE, EISCAT, and opticalobservations, Geophys. Res. Lett., 17, 2241±2244, 1990.

Elphinstone, R. D., J. S. Murphree, D. J. Hearn, L. L. Cogger, P. T.Newell, and H. Vo, Viking observations of the UV daysideaurora and their relation to DMSP particle boundary de®ni-tions, Ann. Geophys., 10, 815, 1992.

Fasel, G. J., J. I. Minow, R. W. Smith, C. S. Deehr, and L. C. Lee,Multiple brightenings of transient dayside auroral forms duringoval expansions, Geophys. Res. Lett., 19, 2429, 1992.

Fasel, G. J., Dayside poleward moving auroral forms: A statisticalstudy, J. Geophys. Res., 100, 11891, 1995.

Feldstein, Y. I., and G. V. Starkov, Dynamics of auroral belt andpolar geomagnetic disturbances, Planet. Space Sci., 15, 209±230, 1967.

Friis-Christensen, E., M. A. McHenry, C. R. Clauer, and S.Vennerstrom, Ionospheric travelling convection vortices ob-served near the polar cleft: a triggered response to suddenchanges in the solar wind, Geophys. Res. Lett., 15, 253, 1988.

Goertz, C. K., E. Neilsen, A. Korth, K.-H. Glassmeier, C. Haldoupis,P. Hoeg, and D. Hayward, Observations of a possible signatureof ¯ux transfer events, J. Geophys. Res., 90, 4069, 1985.

Greenwald, R. A., K. B. Baker, J. R. Dudeney, M. Pinnock, T. B.Jones, E. C. Thomas, J.-P. Villain, J.-C. Cerisier, C. Senior, C.Hanuise, R. D. Hunsucker, G. Sofko, J. Koehler, E. Nielsen, R.Pellinen, A. D. M. Walker, N. Sato, and H. Yamagishi, DARN/SuperDARN: A global view of the dynamics of high-latitudeconvection, Space Sci. Rev., 71, 761±796, 1995.

Haerendel, G., G. Paschmann, N. Sckopke, H. Rosenbauer, and P.C. Hedgecock, The frontside boundary layer of the magneto-pause and the problem of reconnection, J. Geophys. Res., 83,3195, 1978.

Holzworth, R. H., and C.-I. Meng, Mathematical representation ofthe auroral oval, Geophys. Res. Lett., 2, 377±380, 1975.

Horwitz, J. L., and S.-I. Akasofu, The response of the daysideaurora to sharp northward and southward transitions of theinterplanetary magnetic ®eld and to magnetospheric substorms,J. Geophys. Res., 82, 2723, 1977.

Jùrgensen, T. S., E. Friis-Christensen, and J. Wilhjelm, Interplan-etary magnetic ®eld direction and high latitude ionosphericcurrents, J. Geophys. Res., 77, 1976, 1972.

Karlson, K. A., M. éieroset, J. Moen, and P. E. Sandholt, Astatistical study of ¯ux transfer event signatures in the daysideaurora: The IMF By-related prenoon-postnoon asymmetry,J. Geophys. Res., 101, 59±68, 1996.

Lee, L. C., Magnetic ¯ux transfer at the Earth's magnetopause, inSolar Wind-Magnetosphere Coupling, eds. Y Kamide and J ASlavin, Terra Sci., Tokyo, 1986.

Lockwood, M., P. E. Sandholt, and S. W. H. Cowley, Daysideauroral activity and magnetic ¯ux transfer from the solar wind,Geophys. Res. Lett., 16, 33, 1989a.

Lockwood, M., P. E. Sandholt, S. W. H. Cowley, and T. Oguti,Interplanetary magnetic ®eld control of dayside auroral activityand the transfer of momentum across the dayside magneto-pause, Planet. Space Sci., 37, 1347, 1989b.

Lockwood, M., S. W. H. Cowley, P. E. Sandholt, and R. D.Lepping, The ionospheric signatures of ¯ux transfer events andsolar wind dynamic pressure changes, J. Geophys. Res., 95,17113, 1990.

Lockwood, M., The excitation of ionospheric convection, J. Atmos.Terr. Phys., 53, 177±199, 1991.

Lockwood, M., and S. W. H. Cowley, Ionospheric convection andthe substorm cycle, Proceedings of the First InternationalConference on Substorms (ICS-1), Kiruna, Sweden, March1992 (ESA SP-335), 99±110, 1992.

Lockwood, M., and M. N. Wild, On the quasi-periodic nature ofmagnetopause ¯ux transfer events, J. Geophys. Res., 98, 5935±5940, 1993.

Lockwood, M., H. C. Carlson, and P. E. Sandholt, The implicationsof the altitude of transient 630 nm dayside auroral emissions,J. Geophys. Res., 98, 15571, 1993a.

Lockwood, M., W. F. Denig, A. D. Farmer, V. N. Davda, S. W. H.Cowley, and H. LuÈ hr, Ionospheric signatures of pulsed magneticreconnection at the Earth's magnetopause, Nature, 361 (6411),424, 1993b.

Lockwood, M., J. Moen, S. W. H. Cowley, A. D. Farmer, U. P.Lùvhaug, H. LuÈ hr, and V. N. Davda, Variability of daysideconvection and motions of the cusp/cleft aurora, Geophys. Res.Lett., 20, 1011, 1993c.

Lockwood, M., and C. J. Davis, On the longitudinal extent ofmagnetopause reconnection pulses, Ann. Geophys., 14, 865±878,1996.

Milan, S. E., T. B. Jones, T. R. Robinson, E. C. Thomas, and T. K.Yeoman, Interferometer evidence for the observation of groundbackscatter from behind the CUTLASS radars, Ann. Geophys,15, 29±39, 1997.

Moen, J., P. E. Sandholt, M. Lockwood, W. F. Denig, U. P.Lùvhaug, B. Lybekk, A. Egeland, D. Opsvik, and E. Friis-Christensen, Events of enhanced convection and related daysideauroral activity, J. Geophys. Res., 100, 23917±23934, 1995.

Moen, J., M. Lockwood, P. E. Sandholt, U. P. Lùvhaug, W. F.Denig, A. P. van Eyken, and A. Egeland, Variability of daysidehigh-latitude convection associated with a sequence of auroraltransients, J. Atmos. Terr. Phys., 58, 85±96, 1996.

Murphree, J. S., L. L. Cogger, C. D. Anger, S. Ismail, and G. G.Shepherd, Large scale 6300AÊ , 5577AÊ , 3914AÊ dayside auroralmorphology, Geophys. Res. Lett., 7, 239, 1980.

Newell, P. T., and C.-I. Meng, The cusp and the cleft/LLBL: Low-altitude identi®cation and statistical local time variation,J. Geophys. Res., 93, 14549, 1988.

éieroset, M., H. LuÈ hr, J. Moen, T. Moretto, and P. E. Sandholt,Dynamical auroral morphology in relation to ionosphericplasma convection and geomagnetic activity: Signatures ofmagnetopause X line dynamics and ¯ux transfer events,J. Geophys. Res., 101, 13275±13292, 1996.

Paschmann, G., G. Haerendel, I. Papmastorakis, N. Sckopke, S. J.Bame, J. T. Gosling, and C. T. Russell, Plasma and magnetic®eld characteristics of magnetic ¯ux transfer events, J. Geophys.Res., 87, 2159, 1982.

Pinnock, M., A. S. Rodger, J. R. Dudeney, K. B. Baker, P. T.Newell, R. A. Greenwald, and M. E. Greenspan, Observations ofan enhanced convection channel in the cusp ionosphere,J. Geophys. Res., 98, 3767±3776, 1993.

Pinnock, M., A. S. Rodger, J. R. Dudeney, F. Rich, and K. B. Baker,High spatial and temporal resolution observations of theionospheric cusps, Ann. Geophys., 13, 919±925, 1995.


Potemra, T. A., Sources of large-scale Birkelend currents, inPhysical Signatures of Magnetospheric Boundary Layer Process-es, 3±27, Kluwer Academic Publishers, The Netherlands, 1994.

Provan, G., T. K. Yeoman, and S. E. Milan, CUTLASS Finlandradar observations of the ionospheric signatures of ¯ux transferevents and the resulting plasma ¯ows, Ann. Geophys., in press,1998.

Rijnbeek, R. P., S. W. H. Cowley, D. J. Southwood, and C. T.Russell, A survey of dayside ¯ux transfer events observed by theISEE-1 and -2 magnetometers, J. Geophys. Res., 89, 786, 1984.

Rodger, A. S., S. B. Mende, T. J. Rosenberg, and K. B. Baker,Simultaneous optical and HF radar observations of theionospheric cusp, Geophys. Res. Lett., 22, 2045, 1995.

Rodger, A. S., and M. Pinnock, The ionospheric response to ¯uxtransfer events: the ®rst few minutes, Ann. Geophys., 15, 685±691, 1997.

Russell, C. T., and R. C. Elphic, Initial ISEE magnetometer results:Magnetopause observations, Space Sci. Rev., 22, 681, 1978.

Sandholt, P. E., A. Egeland, J. A. Holtet, B. Lybekk, K. Svenes, andS. Asheim, Large- and small-scale dynamics of the polar cusp,J. Geophys. Res., 90, 4407, 1985.

Sandholt, P. E., S. Deehr, A. Egeland, B. Lybekk, R. Viereck, andG. J. Romick, Signatures in the dayside aurora of plasmatransfer from the magnetosheath, J. Geophys. Res., 91, 10063±10079, 1986.

Sandholt, P. E., IMF control of the cusp and cleft auroras, Adv.Space Res., 8, 21, 1988.

Sandholt, P. E., B. Lybekk, A. Egeland, B. Jacobson, P. F. Bythrow,and D. A. Hardy, Electrodynamics of the polar cusp ionosphere:A case study, J. Geophys. Res., 94, 6713, 1989.

Sandholt, P. E., M. Lockwood, T. Oguti, S. W. H. Cowley, K. S. C.Freeman, B. Lybekk, A. Egeland, and D. M. Willis, Middayauroal breakup events and related energy and momentumtransfer from the the magnetosheath, J. Geophys. Res., 95,1039, 1990.

Sandholt, P. E., J. Moen, D. Opsvik, W. F. Denig, and W. J. Burke,Auroral event sequence at the dayside polar cap boundary:Signature of time-varying solar wind-magnetosphere-iono-sphere coupling, Adv. Space Res., 13, (4)7±(4)15, 1993.

Sandholt, P. E., C. J. Farrugia, P. Stauning, S. W. H. Cowley, andT. Hansen, Cusp/cleft auroral forms and activities in relation to

ionospheric convection: Responses to speci®c changes in solarwind and interplanetary magnetic ®eld conditions, J. Geophys.Res., 101, 5003±5020, 1996.

Sandholt, P. E., C. J. Farrugia, J. Moen, N. éystein, B. Lybekk, T.Sten, and H. Truls, A classi®cation of dayside auroral forms andactivities as a function of IMF orientation, J. Geophys. Res., inpress, 1998.

Saunders, M. A., C. T. Russell, and N. Sckopke, Flux transferevents: Scale size and interior structure, Geophys. Res. Lett., 11,131, 1984.

Saunders, M. A., Origin of the cusp Birkland currents, Geophys.Res. Lett., 16, 151, 1989.

Scholer, M., Magnetic ¯ux transfer at the magnetopause based onsingle X-line bursty reconnection, Geophys. Res. Lett., 15, 291,1988.

Siscoe, G. L., and T. S. Huang, Polar cap in¯ation and de¯ation,J. Geophys. Res., 90, 543, 1985.

Smith, M. F., and M. Lockwood, Earth's magnetospheric cusps,Rev. Geophys., 34, 233±260, 1996.

Southwood, D. J., Theoretical aspects of ionosphere-magneto-sphere-solar wind coupling, Adv. Space Res., 5, 7, 1985.

Southwood, D. J., The ionospheric signature of ¯ux transfer events,J. Geophys. Res., 92, 3207, 1987.

Southwood, D. J., C. J. Farrugia, and M. A. Saunders, What are¯ux transfer events?, Planet. Space Sci., 36, 503±508, 1988.

van Eyken, A. P., H. Rishbeth, D. M. Willis, and S. W. H. Cowley,Initial observations of plasma convection at invarient latitudes70±77°, J. Atmos. Terr. Phys., 46, 635, 1984.

Vorobjev, V. G., G. Gustafsson, G. V. Starkov, Y. I. Feldstein, andN. F. Shevnina, Dynamics of day and night aurora duringsubstorms, Planet. Space Sci., 23, 269, 1975.

Walker, A. D. M., R. A. Greenwald, and K. B. Baker, HF radarobservations of pulsations near the magnetospheric cusp,J. Geophys. Res., 91, 8919±8928, 1986.

Wei, C. Q., and L. C. Lee, Ground magnetic signatures of movingelongated plasma clouds, J. Geophys. Res., 95, 2405, 1990.

Whalen, J. A., J. Buchau, and R. A. Wagner, Airbourne iono-spheric and optical observations of noontime aurora, J. Atmos.Terr. Phys., 33, 661, 1971.

Note added in proof Please note that the zenith anglesindicated in Fig. 1 are in the reverse sense to the zenithangle scale of Fig. 3.


meridian-scanning photometer, coherent hf radar, and ......coherent hf radar, a meridian-scanning...

Documents